TABLE 3

Prediction of OS Using Sex, IPI, Stage, Visual Analysis, and SUV-Based Semiquantitative Assessment

Type of assessment	Cutoff	Sensitivity	Specificity	PPV	NPV	Accuracy	2-y survival	P^*
Sex	Female	75%	63%	50%	83%	67%	45%	0.076
	Male						83%
IPI	low = 0, 1	88%	56%	50%	90%	67%	86%	0.047
	high = 2–5						50%
Ann Arbor stage	low = I/II	75%	63%	50%	83%	67%	80%	0.070
	high = IV						50%
DS at midtreatment	low = 1–3	63%	94%	83%	83%	83%	81%	<0.001
	high = 4, 5						17%
SUV_max at pretreatment	low < 7.55	50%	50%	33%	67%	50%	67%	0.965
	high > 7.55						65%
SUV_max at midtreatment	low < 3	63%	94%	83%	83%	83%	81%	<0.001
	high > 3						17%
ΔSUV_max (pretreatment compared with midtreatment)	>88%	88%	13%	33%	67%	38%	50%	0.999
	<88%						67%
DS at end-treatment	low = 1–3	67%	100%	100%	93%	94%	91%	<0.001
	high = 4, 5						0%
SUV_max at end-treatment	low < 2	33%	93%	50%	87%	82%	84%	0.045
	high > 2						50%
ΔSUV_max (pretreatment compared with end-treatment)	>88%	100%	36%	25%	100%	47%	100%	0.246
	<88%						71%